Yambo Community Forum

Davide Sangalli wrote:Thanks for the report.
Did you try to just distribute over "t" ?

Best,
D.

Yes I tried eh16-t4.

As this system has 4 k-points, the maximum t is (4*eh)*(4*eh+1). So with 64 processors, is the model with maximum t k1-eh2-t32? Thanks.

Best,
Feng

There are 4 points in the IBZ, but 6 in the BZ.
The maximum should be 6*eh*7*eh/2=21*eh^2

So, yeah, you can do 2*32.
Some unbalance is expected since the code will distribute 84 transition blocks on 32 CPUs.

Best,
D.